A Hybrid Explanations Framework for Collaborative Filtering Recommender Systems Shay Ben-Elazar Noam Koenigstein Microsoft R&D, Israel { shaybe, noamko }@microsoft.com ABSTRACT Augmenting personalized recommendations with explana- tions is believed to improve users’ trust, loyalty, satisfac- tion, and recommender’s persuasiveness. We present a flexi- ble explanations framework for collaborative filtering recom- mender systems. Our algorithms utilizes item tags to auto- matically generate personalized explanations in a natural language format. Given a specific user and a recommended item, the algorithm utilizes the user’s personal information as well as global information (e.g., item similarities, meta- data) in order to rank item tags based on their “explanatory power”. The top tags are chosen to construct a personalized explanation sentence which helps shed light on the under- lying recommender. Our system has been well received by both focus groups as well as in expert evaluations and is scheduled to be evaluated in an online experiment. 1. BACKGROUND AND INTRODUCTION In the context of recommender systems, an explanation is a short textual description accompanying each recom- mended item to explain the recommender’s choice for the user. The importance of such explanations for real world applications is increasingly recognized. Good explanations are believed to improve user trust, loyalty, satisfaction, and persuasiveness [6]. Experimental evidence has shown that explanations can improve users’ acceptance [1]. The problem of explaining recommendations is not new. Several algorithms have been presented in the past and ex- tensive surveys can be found in [3, 6]. Herlocker et al., [1] distinguishes between “white box’ ’ and “black box ” explana- tion algorithms. For example, in neighborhood models [5] explanations can naturally follow from the recommendation algorithms: ”We recommend Shrek because you watched Toy Story” . This is a “white box ” approach in which the ex- planations are based directly on the recommendation algo- rithm. However, explanations based on similarities are often simplistic and not very engaging. Furthermore, in matrix factorization based systems such as the Xbox recommender Copyright is held by the author/owner(s). RecSys 2014 Poster Proceedings, October 6 - 10, 2014, Foster City, Silicon Valley, USA. system [2, 4], recommendations are based on a global op- timization process and there is no clear method to asso- ciate an explanation item with the recommended item. We therefore turn to “black box ” approaches which justify rec- ommendations in ways which are related to the underlaying recommender but in an indirect manner. In this work we present a “black box ” explanations frame- work based on tags (e.g., as in [7]) in order to compute personalized explanations in the form of “We recommend Shrek based on your history of watching humorous computer animated films featuring imaginary storyline fitting a family outing ”. 2. ALGORITHMIC OVERVIEW Given a recommended item and its tags, we select a small subset of tags that serve best as explanations. These tags are then utilized to automatically generate the explanation sentence. Hence, explanations are computed in two stages: First, we compute scores for the candidate tags to quantify their ability to serve as explanations (“explanatory power”). Then, the tags are ranked and the top tags are chosen to constitute the explanation sentence by applying a predefined rule-set on tag categories. The first stage integrates several modules which constitute scores for either the amount of personalization, quality of the explanation, or quality of the available tags in the context of the recommendation. Our framework allows plugging dif- ferent modules and tuning their relative contribution with corresponding module weights. Each module is represented as a vector or a matrix and can be of one of the following types: a) Tag modules - vectors f k ; b) Item modules -vectors ri ; c) Items to tags relationship modules - matrices Pj ; Our frameworks allows for multiple modules per each type above and we denote by k, i, j the indexes of the modules per type. The scoring algorithm applies an ensemble approach which treats each module as an estimator and computes the weighted average. Let m, n be the number of tags and items, accord- ingly. Then, s m×1 =   i wi r 1×n i  ×   j wj P n×m j  ·   k w k f m×1 k  , (1) and s is the resulting scores vector 1 . Namely, s(l) is the score quantifying the “explanation power” of tag l (see figure 1). The weights wi ,wj ,w k are used to tune the importance of 1 Note the interchanging use of both the dot product and cross product.